1. Important: Please read our guidelines and tips for new players before posting, particularly if this is your first time. If looking for answers, start with a search of past topics.

Hosting - loading speed Aus v USA

Discussion in 'Talking technology' started by Shoes2u, Aug 11, 2012.

  1. Shoes2u

    Shoes2u New Member

    Messages:
    20
    Likes Received:
    2
    Trophy Points:
    3
    I am looking at moving my ecommerce website to US hosting website for various reason mainly PCI compliance.

    Is there a noticeable speed difference in page loading time etc just from being further away? With ecommerce every extra second allows customer time to change their mind so page loading time is very important.

    Thanks
  2. Divert To Mobile

    Divert To Mobile Well-Known Member

    Messages:
    2,633
    Likes Received:
    1,199
    Trophy Points:
    143
    Site load time is also determined by the server load and disk I/O speed.
    Ive seen servers in aus take 15 seconds to send a page and servers in the US come down in 2. (similar content)
    Also if the files in your site a large, Hi res pics - long flash vids etc
    The difference will be more noticable. But the average site will not be as noticable.

    Steve
  3. Greg_M

    Greg_M Well-Known Member

    Messages:
    1,102
    Likes Received:
    694
    Trophy Points:
    113
    I've found it's more about the quality and type of hosting rather than location.

    Shared hosting can be ordinary anywhere, I found a huge difference going to VPS hosting (Linode), sites ran like greased lightning, got sick of managing the server though.

    Currently I'm using a managed platform based in New York, seems fine but I'm only running a couple of light content, low traffic sites so can't really tell, seems quick enough when using the admin backend.

    A good VPS plan wherever it is, should offer a guarantee of available RAM.
  4. John Debrincat

    John Debrincat Well-Known Member

    Messages:
    621
    Likes Received:
    288
    Trophy Points:
    63
    Seems like you are doing exactly the opposite that we see in general in the Australian market. Are you moving the same website that you have now to the USA or changing the technology and rebuilding? If the issue you are trying to solve is PCI compliance and you are not changing the website then just moving it to a different host won't necessarily solve the problem. The host that you use is only part of the factors in the PCI compliance. Running a non-compliant website on a compliant host still means it is non-compliant. Australia has a number of good PCI compliant webhosts that are controlled by Australian regulation and laws and not US. That means you might actually have a chance of pursuing any potential issues if they come up. Your clients data is also protected by Australian law. You can loose all control of your own and client data when hosted overseas.

    The time difference between various locations is referred to as latency. On average between Australia and the USA it is 200 - 250 ms. By comparison between Sydney and Melbourne it is around 30 - 40ms so basically the USA has 5 times more latency. That does not mean that your website slows down by 0.2 of a second. The visitors browser may need to call on the webserver many times 10's even 100's to completely load a website. So the same latency is added for each additional call and is compounding. You can find some information about the latency at http://www.verizonbusiness.com/au/about/network/latency/. However most providers have similar stats.

    If you are hosting where you know there might be some latency then a good web cache strategy is important. You can also investigate content delivery systems (CDN) like CloudFlare or Akamai.

    In addition to the latency there is also the issue of service times and maintenance windows. Service providers in the USA will do maintenance basically in the middle of the Australian business day.

    There is also some impact on SEO but if you use the Google webmaster tool and setup correctly you can over come that to the most part. However if your target business area is Australia then sites of an equivalent rank hosted in Australia will get better local Australian results. If you are targeting your business outside of Australia it will make little difference.

    The other issue is that if you are using shared hosting in the USA you will in general find that there will be more contention for server resources as they seem to pack their systems more densely, at least the cheaper ones do. So if possible ask for service level guarantees that include server resources.

    The key issue really is where your online business is targeted. If it is the USA then hosting there makes a lot of sense. If your core business focus is in Australia then it does not make sense.

    John
  5. Shoes2u

    Shoes2u New Member

    Messages:
    20
    Likes Received:
    2
    Trophy Points:
    3
    It is an x- cart shopping cart and I wanted to change the technology to use x- payments. Ideally I need hosting with good experience in x-cart carts that is also PCI compliant - these seem to only be found outside Australia.

    Core business is catering to Australian market.
  6. Tiggerito

    Tiggerito Active Member

    Messages:
    301
    Likes Received:
    135
    Trophy Points:
    43
    Good stuff John but I have to disagree that server location has an effect on ranking. I believe this is a common misunderstanding based around the fact that Google places every website in a region and for .com domains it uses the server location to work that out. This regionalisation is used when people use the "only show results in xxxx" option which is a filter and has no influence on actual ranking. Local ranking is mainly determined by your content and if local websites link to you.
  7. Zava Design

    Zava Design Well-Known Member

    Messages:
    1,089
    Likes Received:
    368
    Trophy Points:
    83
    Yep, correct.

    And with broadband connection speeds nowadays, for majority of sites there will be no noticeable difference in load speeds.
    1 person likes this.
  8. John Debrincat

    John Debrincat Well-Known Member

    Messages:
    621
    Likes Received:
    288
    Trophy Points:
    63
    @Tiggerito and @Zava Design this issue has been long discussed and here is the Google answer . Basically it is what I wrote. I said that "sites of an equivalent rank hosted in Australia will get better local Australian results.". This is correct and we have seen this happen with sites that we host here in Australia that are targeted to the USA and also overseas targeted at Australia.

    Filtering search results to a particular location is an entirely different issue. What Google has been doing more proactively is serving up local results and you can improve the targeting for local search if you use the right approach - Google local search.

    @Fishx99 you do realise that by using x-payments you just add another layer to the payment process. If you are using a payment gateway why don't you investigate a hosted payment page where the actual payment page is not on your server (like PayPal) but using a local payment provider like eWAY Shared, Payment Express or SecurePay. This will also be PCI compliant as long as the payment page is not on your website.
  9. Shoes2u

    Shoes2u New Member

    Messages:
    20
    Likes Received:
    2
    Trophy Points:
    3
    If I am already using Paypal as hosted payment provider what advantages would there be in offering a local one as well? WIth X-payments my understanding is it is on page payment processing (hence less drop out rate from customers) which is why I was considering going down that path

    Shona
  10. Zava Design

    Zava Design Well-Known Member

    Messages:
    1,089
    Likes Received:
    368
    Trophy Points:
    83
    Yes, and as that same page says, it will first check webmaster tools, which is the first stop for any new site going live (or should be), and only then it will rely on a range of other factors, of which IP is one of them, but no information given about the weighting against the many other factors they also include. And as I said, my tests have shown negligible IP address weighting compared to the range of other factors.

    And I for one am certainly not going to advise my clients to pay more for less with regards to web hosting with negligible real world effect.

    Now you have another opinion, that's fine, but I would argue mine is based on just as careful research as yours, we've just come to different conclusions. Won't be the first time that's every occurred, nor the last.
    Last edited: Aug 13, 2012
  11. John Debrincat

    John Debrincat Well-Known Member

    Messages:
    621
    Likes Received:
    288
    Trophy Points:
    63
    No expert on x-payments but my understanding is that the payment page is in x-payments and not your website. You install x-payments and then configure a connection to your website. You have to create a skin for the payment page in x-payments to look like your pages. Which is why, I guess, it is PCI compliant because the x-payment application is PA-DSS, so although it looks the same as your website it is not a page on your website. You will find that some of the hosted payment solutions can do a pretty nice job of keeping the on-page approach using ajax or java script to automate the payment page.
    1 person likes this.
  12. John Debrincat

    John Debrincat Well-Known Member

    Messages:
    621
    Likes Received:
    288
    Trophy Points:
    63
    What you advise your clients is always entirely up to you. But just do a simple ping compare between your website hosted in the USA and ours hosted in Australia. The average is 224ms versus 29ms. For a simple website that does not mean a lot as you say but for an online store driven by a data base it will mean a lot of difference in load time.

    Attached Files:

  13. MatthewKeath

    MatthewKeath Well-Known Member

    Messages:
    2,750
    Likes Received:
    1,386
    Trophy Points:
    143
    As others have mentioned, most hosted payment processors have an option of processing payments directly on your website.

    Paypal has one as well - if you are already using them why not look into it?
  14. Zava Design

    Zava Design Well-Known Member

    Messages:
    1,089
    Likes Received:
    368
    Trophy Points:
    83
    If you have a badly optimised site perhaps. However a well built, well optimised site, the different between those milliseconds when you're on a current average broadband connection is negligible.

    As I said previously, for sites that may be heavy on the video content, that would be the only type of site that a user may notice a difference, and even then only on occasions.
    Last edited: Aug 13, 2012
  15. John Debrincat

    John Debrincat Well-Known Member

    Messages:
    621
    Likes Received:
    288
    Trophy Points:
    63
    OK let me try once more and then I give up as it seems like a waste of time trying to explain this. Try using this web resource and test your own website from different locations - http://www.webpagetest.org/. Your website is hosted in Dallas Texas and the nearest test server was San Jose. Your site total load time was 4.813 seconds and first connection was 0.512 secs. When tested from Singapore (similar latency to Australia / USA) the load time was 9.762 secs and first connection was 1.893 secs. So about twice as fast if you are in the USA.

    I don't see anything special in the way of content on your website.

    Latency is defined in milliseconds (ms) but it is not a one time addition to the page load time and it is not the only influencing factor.

    You can also try http://tools.pingdom.com/fpt/ they have a Dallas and Amsterdam test server and you will see that from Amsterdam it is about double the load time. Latency between Dallas and Amsterdam is a lot less that Sydney / Dallas.

    So if a customer asks you will their website take twice as long to load in Australia if it is hosted in the USA would your advice be that "it does not matter"?
    1 person likes this.
  16. Zava Design

    Zava Design Well-Known Member

    Messages:
    1,089
    Likes Received:
    368
    Trophy Points:
    83
    Then should I as well? I understand exactly what you're saying, you seem to be missing mine.

    Difference is milliseconds.
    On average Australian broadband speed this difference is negligible.

    Is that clear now?

    And as a result of that negligible difference, I advise clients (or even those who aren't clients) to spend time & budget on more important areas, such as their actual website, the quality of their hosting, copywriting, marketing ....etc.
    Last edited: Aug 13, 2012
  17. John Debrincat

    John Debrincat Well-Known Member

    Messages:
    621
    Likes Received:
    288
    Trophy Points:
    63
    The difference in your own website is 4 - 5 seconds not milliseconds. I would say that was a significant difference for your simple website. Now consider that for a complex or more content rich online store.
  18. Zava Design

    Zava Design Well-Known Member

    Messages:
    1,089
    Likes Received:
    368
    Trophy Points:
    83
    I don't rely on virtual tests, I have done many real world tests over the years, both for small and large sites.
  19. Shoes2u

    Shoes2u New Member

    Messages:
    20
    Likes Received:
    2
    Trophy Points:
    3
    If I use PayPal on page processing then I am back to my issue again of needing to find hosting that is PCI compliant. I want hosting that specifically deals with xcart software to nullify any issues. Not many options for that in Australia other than using a dedicated server which is too expensive.
  20. Tiggerito

    Tiggerito Active Member

    Messages:
    301
    Likes Received:
    135
    Trophy Points:
    43
    I tend to recommend an Australian host. Partly for the potential speed improvement and also to help support local business.

    I do not believe local hosting has any influence on ranking though (apart from extreme cases of very slow sites).

    The cited article from Google on geo targeting in Webmaster Tools specifically states it has no effect unless a user "limits the scope of a search to a particular country". i.e. they imply it's a filter and not part of the ranking algos.

    My own limited experience with geo targeting international websites has not show any clear ranking changes when geo targeting has been changed.

    There's also the debate on using ccTLDs. I recommend their use for any national website but more because users prefer them. The jury is still out on their effect on ranking.
    1 person likes this.

Share This Page